The Tournalyzer simulates hold'em tournament situations where the players are in "all-in or fold" mode before the flop, as typically occurs when the blinds are high and there are few players left.

This is best answered by example.

Let's assume we are at the end of a ten-player $10 no-limit "sit-n-go" tournament, with prizes of $50, $30, and $20. Four players remain, with blinds of 50 and 100. The under-the-gun player raises all-in for 1000 chips. Given our observations of the all-in raiser, we believe he could go in with a very wide range of hands here - as an estimate, we put him on any pair, any bigger than average ace, any suited ace, any king-queen, and king-jack suited. The next player folds with 2000 chips. The small blind folds with 4950 chips (after posting 50). We are last to act in the big blind with 1900 chips after posting the big-blind of 100. We are dealt a pair of eights, and decide to go all-in.

Click here to see this situation in the hold'em tournalyzer editor. Everything should be fairly self explanatory.

Go ahead and click the "Tournalyze" button now. The Tournalyzer will perform a (very) large number of trials. In each trial, it will deal cards to each player, with the guarantee that anyone with "force" checked gets dealt a hand from the set specified. After all the hands are dealt, a five card board is generated. Finally, the tournalyzer figures out who wins the hand, how many chips each player has at the end of the hand, and how much money each player expects to win in the tournament given the size of their stack at the end of the hand.

The pre-hand values show you the values of each player's stack before the simulation was run. The post-hand values show you the average value of each player's stack after the simulation. If a player has a green post-hand value, then that player stands to make money on average after this hand, whereas a red hand is a money loser, and a black hand is close to even. Note that in this example, the big winners were the players who folded their hands.

It is important to note that we can't necessarily judge the quality of a play by whether it yields a red, green or black post-hand value. In this example, the first player expects to lose money on average, but that is only because we know he is getting called by a pair of eights - there is no way he could have known that at the time. Furthermore, situations can occur where all plays for a certain player are in the red - this typically occurs when a player is in the blind facing a raise. In this case, the goal is to find the play that loses the least.

If "Always" is selected for a given player, then every trial will result in that player being dealt a hand from the list specified. If "Conditional" is selected, then that player will fold if he is not dealt a hand from the list.

Sure thing. Let's say we are at the beginning of a $50 six-player tournament with payouts of $200 and $100. You and the first player play with each other all the time, and you know for a fact that he always goes all-in on the first hand with any two cards (if he is the first one in), and shows his hand afterwards for dramatic purposes. You are next to act with T9s - let's go ahead and call all-in with it and see what happens. All the other players are reasonably tight - we assume they won't over-call all-in without a monster - let's say AA-QQ for now.

Click here to see this situation in the hold'em tournalyzer editor. We choose "Always" all-in for the first two hands because we know these two players are going all-in. The other players may or may not go all-in depending on what they are dealt, so we choose "Conditional". If you run the simulation, you'll find that calling is a slight loser - note that if you were last to act in the big-blind after everyone had folded, calling would be a small winner (Give it a try if you like).

As is well known in poker tournament circles, the value of a chip is not constant in a typical tournament. For instance, in our first example, ten players entered the tournament for $10 and received 1000 chips. At this point, it should be clear that each chip is worth one penny. At the end of the tournament, one player will have 10,000 chips and claim a prize of $50. But this implies the value of a chip has gone down to half of a penny.

The ProPokerTools Tournalyzer uses a popular computational method for estimating the value of a stack given the size of all the other stacks and the payout schedule. It is known as the Independent Chip Model, and is somewhat beyond the scope of our discussion here - an explanation with links to more information can be found here.

When every possible combination of hands and boards are computed, we say the results are exhaustive. When a large number of randomized results are run, we say the results are random. Exhaustive results will be the same every time, while random results will differ slightly from run to run. Curently, we provide exhaustive results for simulations involving two players.

Exhaustive results are provided when there are two players involved in the hand, regardless of the number of players who are seated in the tournament.

Excellent question! A simulator is next to useless if you don't understand exactly what it is doing. In no particular order:

• In the event that all players fold, the last player listed gets all of the blinds. For this reason, you should list players in order of action from top to bottom, with the big blind last.
• In the event that two players bust out at the same time, the player with the larger stack at the start of the hand gets placed higher. If the chip counts are the same, it is considered a tie. For instance, if we are down to four players where third pays $20 and fourth pays nothing, and two players bust out, the player with the larger stack at the start of the hand gets $20 - in the event they have the same stack size, they both receive $10.
• No fractional chips are used - in the event that chips cannot be divided evenly, the players receive extra chips from top to bottom until none remain. For example, in a three way split with 14 chips in the pot, player one gets five chips, player two gets five chips, and player three gets four chips. Using this method each player may receive at most one extra chip from any split pot.

This list may not be complete - if you have a question about something, please use the contact link at the bottom of the page to let us know.

This should only happen in a trivial simulation (for example, where one player always goes all-in and nobody calls).

What an excellent idea! This is on our to-do list.

A fair question. It involves a lot of automated testing, in additon to a fair amount of low tech investigation of log files. If you find a suspicious result, please send the link to the results page using the Contact link at the bottom of this page.

Well, there are a number of possibilities that come to mind:

• Our tool has a bug
• Their tool has a bug
• Our tools have different assumptions
• You did not actually enter the same simulation
• You are comparing randomized simulations, which, by their very nature, will never be exactly equal

If you are truly stumped, send us all the information using the contact page below and we'll see if we can figure it out.